Page 1
Instructions for use
Title Phase-Separation-Induced Anomalous Stiffening, Toughening, and Self-Healing of Polyacrylamide Gels
Author(s) Sato, Koshiro; Nakajima, Tasuku; Hisamatsu, Toshiyuki; Nonoyama, Takayuki; Kurokawa, Takayuki; Gong, Jian Ping
Citation Advanced Materials, 27(43), 6990-6998https://doi.org/10.1002/adma.201502967
Issue Date 2015-11-18
Doc URL http://hdl.handle.net/2115/65109
Rights
This is the peer reviewed version of the following article: Sato, K., Nakajima, T., Hisamatsu, T., Nonoyama, T.,Kurokawa, T. and Gong, J. P. (2015), Phase-Separation-Induced Anomalous Stiffening, Toughening, and Self-Healingof Polyacrylamide Gels. Adv. Mater., 27: 6990‒6998, which has been published in final form athttp://dx.doi.org/doi:10.1002/adma.201502967. This article may be used for non-commercial purposes in accordancewith Wiley Terms and Conditions for Self-Archiving.
Type article (author version)
Additional Information There are other files related to this item in HUSCAP. Check the above URL.
File Information ksato_advmat_compmanu-rev.pdf
Hokkaido University Collection of Scholarly and Academic Papers : HUSCAP
Page 2
1.
1
DOI: 10.1002/((please add manuscript number)) 1
Article type: Communication 2
3
4
Phase Separation-Induced Anomalous Stiffening, Toughening, and Self-healing of 5
Polyacrylamide Gels 6
7
Koshiro Sato, Tasuku Nakajima, Toshiyuki Hisamatsu, Takayuki Nonoyama, Takayuki 8
Kurokawa, Jian Ping Gong* 9
10
K. Sato, T. Hisamatsu 11
Laboratory of Soft & Wet Matter, Graduate School of Life Science, Hokkaido University, 12
Kita-10, Nishi-8, Kita-ku, Sapporo, Hokkaido, 060-0810, Japan 13
E-mail: [email protected] 14
15
Dr. T. Nakajima, Dr. T. Nonoyama, Dr. T. Kurokawa, Prof. J. P. Gong 16
Laboratory of Soft & Wet Matter, Faculty of Advanced Life Science, Hokkaido University, 17
Kita-10, Nishi-8, Kita-ku, Sapporo, Hokkaido, 060-0810, Japan 18
19
Keywords: polymer gel, phase separation, toughening, sacrificial bond, self-healing 20
21
Page 3
1.
2
1
Abstract 2
Mechanical properties of the common polyacrylamide (PAAm) hydrogels are dramatically 3
improved simply by inducing appropriate phase separation structure using poor solvent. By 4
immersing the PAAm hydrogels into the solvents with various affinities, the gels shrank with 5
decrease of affinity, and at certain point the gels started to show micro phase separation. 6
Around the phase separation point, the gels exhibited anomalously high stiffness (modulus: 7
211 MPa), strength (fracture stress: 7.13 MPa), extensibility (fracture strain: 535 %), and 8
toughness (fracture energy: 4.16 × 104 J m
-2) while keeping high solvent content of ~60 vol%. 9
The fracture energy observed in this sample is among the highest value ever obtained in gel 10
systems. The gel also showed full self-healing up to a strain of 500%. The tremendous 11
increase of the toughness of the phase-separated gels could be understood by the double 12
network effect; that is, the aggregated polymers in the dense phase serve as reversible 13
sacrificial bonds to dissipate energy at deformation, which imparts the strength, toughness, 14
and self-healing of the material. 15
16
Introduction 17
Polymer gels are cross-linked polymer network systems containing plenty of solvents. The 18
characteristics of gels, such as optical, mechanical, and thermal properties, are strongly 19
related to their internal structure. The latter sensitively changes with the solvent affinity to the 20
polymer. In good solvents, a gel swells extensively, showing high transparency and elastic 21
property. In poor solvents, a gel shrinks, usually becoming opaque due to micro phase 22
separation. Atomic force microscopy (AFM) and laser scattering confocal microscopy studies 23
have revealed that the structure of phase separated gels is composed of dense and sparse 24
polymer regions in real space and they form sub-micrometer scale bicontinuous domains.[1,2]
25
Although much attention has been paid to understanding the structure of the phase separated 26
Page 4
1.
3
polymer gels, until now few attentions have been paid to the mechanical properties. 1
Shibayama et al. investigated the relationship between phase separation of poly(N-2
isopropylacrylamide) gels and their rheological behaviors, and found the viscoelastic property 3
of phase separated gels.[3]
However, the mechanical performances of phase separation gels, 4
such as fracture strength and fracture energy, have not investigated well in spite of the 5
importance in applications. 6
The aim of this study is to investigate the tensile deformation behavior and fracture behavior 7
of phase separated polymer gels. In a good solvent, a gel is purely rubber-elastic, with 8
negligible interaction between polymers. The fracture energy of an elastic gel is relatively low, 9
as well described by the Lake-Thomas theory.[4]
That is, the energy to fracture a unit surface 10
equals to the areal density of the polymer chains across the fracture surface multiply the total 11
primary bond dissociation energy of each chain. This theory predicts the fracture energy of 12
typical gels in the order of 100~10
1 J m
-2, which is well in agreement with experimental 13
observation. [5, 6]
In a poor solvent or when a polymer volume fraction is very high, however, 14
the interaction between the polymers, either directly or via solvent, becomes important. 15
Masuike et al. have found that air-dried transparent polyacrylamide gels (PAAm) with high 16
polymer density (about 78vol%) show significant increase of modulus and the loss tangent δ 17
on rheological test, suggesting the presence of strong associative interaction between the 18
polymers.[7]
For a material contains viscous term, the fracture energy increases in proper 19
velocity due to this additional viscous energy dissipation around crack tip.[8]
Thus, formation 20
of the dense polymer domains by phase separation is expected to increase the toughness of 21
polymer gels. Thus, it is extremely meaningful to study “real toughness” of such phase 22
separated gel system. 23
In this study, common PAAm gels were used as model system of phase separated gels. These 24
gels were immersed into mixtures of N,N-dimethylformamide (DMF)/water solutions that are 25
poor/good solvent of PAAm, respectively.[9]
As a comparison, the de-swelling was also 26
Page 5
1.
4
performed by immersing the gels in concentrated poly(ethylene glycol) (PEG, Mw = 20,000) 1
aqueous solutions. The latter method leads to uniform de-swelling of the gels without phase 2
separation by the high osmotic pressure of the PEG solution.[10]
The affinity of solvent and the 3
osmotic pressure of the water bath can be precisely controlled by changing the concentration 4
of poor solvent DMF (CDMF, wt%) and PEG concentration (CPEG, wt%) respectively. The 5
PAAm gels de-swollen in poor DMF/water solutions are denoted as p-PAAm gels, and the 6
gels de-swollen in PEG solutions by osmotic pressure are denoted as o-PAAm gels. The 7
fractions of DMF and PEG are described as a number before p- or o- (for example, 80p-8
PAAm represents the PAAm gels immersed in 80wt% DMF aqueous solution and 40o-PAAm 9
immersed in 40wt% PEG aqueous solution). 10
In this paper, we report the mechanical properties and fracture behavior of the de-swollen gels. 11
Effect of polymer volume fraction ϕ and the de-swelling method on the mechanical properties 12
of gels are discussed in detail. The behaviors between uniformly de-swollen gels and phase-13
separated gels with the same overall polymer volume fraction are compared to clarify the 14
effect of phase separation on the mechanical properties. 15
16
Results and Discussion 17
Appearance Change of Gels 18
Figure 1a, b show the pictures of the p-PAAm gels and the o-PAAm gels after being 19
immersed in DMF/water and in PEG solutions, respectively. In both systems, the volume of 20
the gels remarkably decreased with increase of CDMF and CPEG due to decrease of affinity 21
between solvent and PAAm and increasing in osmotic pressure in the water bath, respectively. 22
Figure 1c, d show polymer volume fraction ϕ of the gels against CDMF or CPEG, respectively. ϕ 23
of the gels increased from 0.03 to about 0.5 in both DMF and PEG systems. ϕ of o-PAAm 24
gels increased linearly with increase of CPEG whereas ϕ of p-PAAm gels showed abrupt 25
increase around 60wt% of CDMF and reached to almost constant value once attaining a critical 26
Page 6
1.
5
CDMF of 80wt%. The o-PAAm gels were completely transparent regardless of CPEG, while p-1
PAAm gels became opaque after immersion with CDMF > 80wt% (see Figure 1a). This result 2
indicates that immersing of PAAm gels in concentrated DMF aqueous solutions causes phase 3
separation due to low affinity of DMF to PAAm. This observation is in well-agreement with 4
the previous study, which reported that linear PAAm shows obvious phase-separation in 5
80wt% DMF aqueous solution.[11]
Note that CDMF was weight concentration of DMF in the 6
bath solvents, not in the gels. The true CDMF in the gels was estimated by NMR measurement 7
and the results are shown in Table S1 of Supplementary Information. 8
9
Tensile Properties of p- and o-PAAm Gels 10
Figure 2a, b show typical tensile stress-strain curves of the p- and o-PAAm gels, respectively. 11
For comparison, the data of the original water-equilibrated PAAm gel is also shown in the 12
figures. The water-equilibrated PAAm gel (ϕ = 0.031) is soft and elastic with low Young’s 13
modulus of 4.58 ± 0.17 kPa, low fracture stress of 32.50 ± 1.84 kPa, and high fracture strain 14
of 6.31 ± 0.17. In the case of p-PAAm gels, when CDMF was less than 80wt%, corresponding 15
to polymer volume fraction ϕ less than 0.43, the gels were also soft and elastic. When CDMF 16
reached 80wt%, the gels became glassy-like, showing abrupt increase in Young’s modulus and 17
fracture stress and decrease in fracture strain. Furthermore, at CDMF = 80wt%, the p-PAAm gel 18
showed clear yielding behavior, like thermoplastic polymeric materials while keeping 60vol% 19
of solvent content. Note that the sample did not show necking phenomenon unlike tough 20
double network gels (Figure S1a,b).[12]
On the other hand, when CDMF was higher than 21
90wt%, the gels turned brittle and showed much lower fracture strain εf, less than 0.1 mm/mm. 22
These samples did not show clear yielding behavior. In contrast, the o-PAAm gels showed 23
small and gradual change in the tensile stress-strain curves with the increase of CPEG. 24
The Young’s modulus E and fracture strain εf of the gels de-swollen in the two different 25
solutions are summarized in Figure 2c,d. The E and εf of o-PAAm gels gradually increased 26
Page 7
1.
6
with the PEG concentration until CPEG = 45wt% (Figure 2c). In the case of the p-PAAm gel 1
systems, E abruptly increased at CDMF = 60wt% and εf abruptly dropped at CDMF = 80wt% 2
(Figure 2d). 3
As shown in Figure 2e, E of the o-PAAm gels against ϕ showed a scaling relationship of 4
𝐸~𝜙0.98 (1) 5
This is consistent with the classical rubber elasticity theory, that is 6
𝐸 = 3𝜙𝜈𝑒𝑅𝑇 (2) 7
where ve is elastically-effective polymer chain density at the solvent free state, R is gas 8
constant and T is absolute temperature.[13]
Thus, the relationship of Eq.1 means that the de-9
swollen o-PAAm gels maintain the elastic property regardless of ϕ at least until ϕ = 0.50. 10
The E of the p-PAAm gels against ϕ in the range ϕ ≤ 0.38 (CDMF ≤ 70wt%) also obeys the 11
same scaling relationship with that of o-PAAm gels, as shown in Figure 2e. This means that 12
the de-swollen p-PAAm gels before abrupt stiffening were also in rubbery state. The abrupt 13
change in mechanical properties of p-PAAm gels around 75wt% DMF can be deemed as gel-14
to-glasslike transition. As shown in Figure 2e, the E of the p-PAAm gels in the region ϕ > 15
0.38 no longer follows the rubber elasticity theory. E of such gels increased abruptly with 16
increasing in ϕ and reached to ~103 MPa, which is 10
4 times higher than that of the o-PAAm 17
gels of the same ϕ. These mechanical test results show two important facts. One is that the 18
mechanical transition of the PAAm gels was not only triggered by change of polymer volume 19
fraction. In the case of o-PAAm gels, which did not show phase-separation, they were still 20
soft and elastic even contained less amount of solvent than that of the phase-separated p-21
PAAm gels. The other is that CDMF at the mechanical transition point (70~80wt%) is very 22
close to critical CDMF for visible phase separation (Figure 1a), which implies that such 23
mechanical transition of the PAAm gels is induced by the phase separation. 24
The solvent-induced phase separation was reversible. Once these phase separated gels were 25
immersed into water and were equilibrated, the phase-separation structure disappears and the 26
Page 8
1.
7
gels recover their original sizes mechanical properties. For example, εf and E of 80p-PAAm 1
after immersion in water were 6.82 ± 0.89 and 5.01 ± 0.21 kPa, while those of the originally-2
water-swollen PAAm were 6.31 ± 0.17 and 4.58 ± 0.17 kPa, respectively (Figure S1c). 3
4
Fracture Energy by Pure Shear Test 5
To evaluate the toughness of p- or o-PAAm gels, we measured the fracture energy T of the 6
samples by the pure-shear test of notched specimens following the previous studies (Figure 7
S2).[14,22]
As shown in Figure 3a, the p-PAAm gels in the transition region (CDMF = 8
70~80wt%) were very tough and showed excellent resistance to rupture. Despite containing 9
60wt% of solvent, these phase-separated tough gels showed fracture energy of ~104 J m
-2, 10
while the maximum T of o-PAAm gels with higher polymer density than phase-separated 11
tough gels was ~103 J m
-2 (Figure 3b). To the best of our knowledge, the maximum T = (4.16 12
± 0.38) × 104 J m
-2 observed for the p-PAAm immersed in 80wt% DMF is the among the 13
highest fracture energy of gels ever observed, especially with a very high modulus (211 MPa). 14
For example, double network (DN) gels, which have been known as extremely tough 15
hydrogels despite 90wt% of water content, show T of 102~10
4 J m
-2, with modulus in sub-16
MPa range.[ 15, 16, 17, 18]
Recently-developed hybrid gels from alginate-PAAm and from PVA-17
PAAm show T of ~2 × 104 J m
-2 with a modulus of 1~5 MPa.
[19, 20] However, the p-PAAm 18
sample immersed in 90wt% DMF became very rigid and brittle, they fractured in making a 19
notch in the test pieces. 20
To demonstrate the toughness and strength of the phase separated p-PAAm gels, we 21
performed leaf wrestling “kusa-sumo” of the p-PAAm gel versus a commercially available 22
silicone rubber that is known as a tough structural material. Kusa-sumo is a traditional 23
Japanese children game, comparing the toughness and strength of two stems or leaves by 24
pulling them toward the opposite direction (Figure 3c). First, both a sheet of silicone rubber 25
and a sheet of 80p-PAAm gel were cut into dumbbell shape with 12 mm length and 2 mm 26
Page 9
1.
8
width. Thickness of the silicone and the gel were 2.00 mm and 1.55 mm respectively. Then 1
the gel sample was annealed in 80wt% DMF solution at 70oC to bend it into a loop, and then 2
was cooled down in the same solvent at 25oC. The silicone sample was then entangled to the 3
loop of gel sample, and the ends of silicone and gel were fixed by clamps of the tensile tester. 4
During the kusa-sumo test, the silicone rubber fractured first although it had an advantage of 5
larger thickness (Figure 3d). This result indicates that by simply immersing the weak PAAm 6
hydrogels into poor solvent, one can make the material stronger than the tough silicone rubber. 7
8
Mechanical Hysteresis Loss of p-PAAm gels in Transition Region 9
The extra-ordinarily high toughness of the p-PAAm gels near the gel-to-glassy transition 10
region is considered to be due to viscous energy dissipation by the aggregated polymers in the 11
dense phase. Loading and unloading tensile test was performed to elucidate the mechanism of 12
high toughness of the gel. The area between loading and unloading stress-strain curves of the 13
tensile test, known as mechanical hysteresis loss Uhys, corresponds to the energy dissipated 14
during deformation of the sample. When the sample shows excess energy dissipation during 15
deformation, it possesses large Uhys and high toughness. Figure 4a, b is the data of loading 16
and unloading tensile test for non-phase separated and phase separated p-PAAm gels at a 17
strain range of 0 ≤ ε ≤ 5, respectively. Whereas non-phase separated gels shows negligible or 18
very small Uhys, 10-3
~10-1
MJ m-3
(Figure 4a), the phase separated gel showed extremely 19
large Uhys, more than 10 MJ m-3
(Figure 4b). The work of extension at fracture Wextf and 20
mechanical hysteresis loss Uhys (0 ≤ ε ≤ 5) of p- and o-PAAm gels are summarized in Figure 21
4c,d. The o-PAAm gels showed monotonous increase in the Wextf and Uhys with the CPEG, and 22
the maximum Wextf was as high as 10 MJ m-3
, while Uhys was quite low (~10-1
MJ m-3
) 23
(Figure 4c). On the other hands, the p-PAAm gels showed a maximum of Wextf as high as 50 24
MJ m-3
at CDMF = 78 wt%. The Uhys of this p-PAAm gel was also very high (~101 MJ m
-3) 25
(Figure 4d). This large energy dissipative property and high toughness of p-PAAm lost once 26
Page 10
1.
9
the gels were equilibrated in water again, and the gels returned back to the initial soft and 1
elastic properties in water (Figure S1c, d). 2
3
Self-healing of Sacrificial Bond 4
From the result of loading and unloading test, phase separated gels showed excess energy 5
dissipation by internal structural change, which describes the toughening mechanism of those 6
gels. To investigate whether the internal structural change is permanent or self-healable, we 7
performed loading and unloading tensile test for one sample repeatedly. Figure 4e, f show 8
repeated tensile loading and unloading (0 ≤ ε ≤ 3) of a 80p-PAAm sample that showed large 9
energy dissipation in deformation. Self-healing was observed even in the room temperature 10
without any thermal annealing (Figure 4e). In the 2nd
and 3rd
tests, in which the waiting time 11
between the subsequent tensile cycles was short (10 min), Uhys were smaller than that of the 12
virgin sample, indicating the partial self-healing. The value of Uhys in the 3rd
test, in which the 13
waiting time between the subsequent tensile cycles was 24 hours, was 97% of Uhys of initial 14
test (Figure 4e). This indicates the sample self-healed fully after 24 hours. To activate the 15
recovery, the sample was thermal annealed between the subsequent tensile cycles by 16
immersing in the corresponding DMF/water solution of 75oC for 5 min, above its softening 17
temperature (Figure 4f). To make a comparison, the virgin sample was also annealed at 75oC 18
for 5 minutes to incorporate the annealing effect on the mechanical properties. By heat to 19
75oC, the storage modulus of the sample decreased from 1.20 × 10
2 MPa to 1.27 × 10
-1 MPa. 20
No opaque-transparent transition was observed by heating at high temperature (Figure S3a). 21
For the three repeated loading and unloading tests, the annealed sample showed almost the 22
same stress-strain curve and the value of hysteresis loss (Uhys = 7.84 ± 0.60 MJ m-3
). These 23
results suggest that the structure change that dissipates energy of p-PAAms is reversible, 24
which serves as the reversible sacrificial bond to toughen the material. Moreover, the self-25
healing rate can be accelerated by raising temperature. 26
Page 11
1.
10
1
Rheological Experiments 2
Following the previous works on phase-separated hydrogels, we also performed rheological 3
experiments of the two kinds of the gels. Figure S3b, c show temperature dependence of the 4
storage modulus G’, the loss modulus G’’, and the loss tangent tan δ of p-PAAm gels. In 5
rubbery region (CDMF < 70wt%, ϕ = 0.031 in S3c), tan δ of the gels was less than 0.1 in all 6
temperature range due to their elastic nature. This is a typical behavior of chemically-cross-7
linked gels with plenty of solvent. In transition region (70wt% ≤ CDMF ≤ 80wt% in S3c), the 8
gels showed large tan δ near room temperature. Surprisingly, the peak tan δ showed quite high 9
value of around 2. It means that the gels in transition region were highly viscoelastic, or even 10
behaved like viscous liquid. With increase of CDMF, peak of tan δ shifted to high temperature. 11
Finally in brittle region (CDMF > 80wt%, ϕ = 0.40 in S3c), tan δ of the gel at room temperature 12
became small again. Combining these rheological results and the large hysteresis loss in 13
loading and unloading tensile test, it is clear that phase-separated structure in gels, inducing 14
large mechanical hysteresis and viscoelasticity, contributes to extremely high toughness of p-15
PAAm gels. In short, we can conclude that the strong polymer-polymer interaction by phase-16
separation serves as reversible sacrificial bonds to impart toughness and self-healing of the p-17
PAAm gel. 18
19
Discussion 20
Previously Masuike et al. has reported that the air-dried PAAm gels showed gel-to-glasslike 21
transition at a critical polymer volume fraction ϕc = 0.78.[7]
This explains why all of the o-22
PAAm gels, dehydrated homogeneously (ϕ < 0.5), show elastic behaviors. On the other hand, 23
the p-PAAm gels show the phase separation and mechanical transition around ϕ = 0.40, which 24
is much lower than ϕc = 0.78. When the phase separation occurs, polymer dense phases and 25
sparse phases are formed in the gels. Thus, although the average ϕ in bulk is still much lower 26
Page 12
1.
11
than ϕc, the local polymer volume fraction in the dense phase reaches the critical value of ϕc = 1
0.78, to cause the gel-to-glasslike transition. This speculation is supported by the value of 2
Young’s modulus of such p-PAAm gels. Young’s moduli of both air-dried PAAm gels and p-3
PAAm ones reached ~100 MPa after mechanical transition. 4
Figure 5 shows the hypothesis of mechanical behaviors in each region. Before phase 5
separation (CDMF < 70wt%), polymer chains are highly solvated so that polymer-polymer 6
interaction in gels is screened. As a result, gels are almost purely elastic with negligible 7
hysteresis loss (viscous energy dissipation), and the elasticity of the gels follows the classical 8
rubber elasticity theory (Figure 4a, 5a). Around the phase separation point (70wt% ≤ CDMF ≤ 9
80wt%), the solvent becomes poor enough that the polymer-polymer association prevails to 10
form local dense phase. When such gels are elongated, the polymer-polymer association in the 11
dense phases are preferentially destroyed during deformation, providing a mechanism as the 12
sacrificial bonds to dissipate energy, in similar to the double network hydrogels.[21]
As a result, 13
the mechanical hysteresis and fracture energy substantially increase. Different from double 14
network hydrogels, the ruptured polymer-polymer associations are reformed during loading, 15
which keeps the deformed structure to give large residual strain in short time scale (Figure 4b, 16
5b). As the polymer chains are chemically crosslinked, the rubber elasticity of the polymer 17
chains provides a driving force to recover the polymer chains to the original conformation. As 18
a result, the gels recover their original length and mechanical properties after the loading, 19
which can be called self-healing of sacrificial bonds (Figure 4e). If annealing is performed 20
above the softening temperature of the glassy-like polymer-polymer association, the recovery 21
rate is accelerated (Figure 4f). On the other hand, when CDMF is higher than 80wt%, the dense 22
phase develops to form stronger polymer-polymer associations, as shown by the abruptly 23
increase of the modulus and the strength of the gels (Figure 2d). As a result, the association 24
of the polymers becomes too strong to rupture as sacrificial bonds and no energy dissipation 25
occurs (Figure 5c). As a result, the gels become brittle at high CDMF. Above 80wt% CDMF, 26
Page 13
1.
12
although the global polymer volume fraction ϕ almost does not change with the further 1
increase of CDMF, the mechanical behavior is different. This is due to difference of internal 2
CDMF. NMR analysis showed that the internal CDMF of 80p-PAAm is 54.6wt% and that of 3
100p-PAAm is 80.5wt%. Therefore, the polymer-polymer interaction of 100p-PAAm is 4
stronger than that of 80p-PAAm. 5
This toughening mechanism of p-PAAm gels is in common to that of recently reported stiff 6
and tough gels such as polyampholyte gels (PA gels) [22]
and PVA-PAAm hybrid gels in which 7
breaking of physical bonds is the mechanism of energy dissipation.[20]
However, the 8
maximum fracture energy of p-PAAm gels reaches to ~ 40,000 J m-2
, which is higher than PA 9
gels (~ 4,000 J m-2
) and PVA-PAAm hybrid gels (~ 14,000 J m-2
) even with a much higher 10
stiffness (~ 200 MPa) than these systems. The difference in the fracture energy is considered 11
to be due to the difference in G’’, which strongly relates to the amount of viscous energy 12
dissipation. Optimized p-PAAm gels show extremely-high G’’ of 108 Pa as shown in Figure 13
S3. On the other hand, G’’ of the PA gel at the same condition is 106 Pa, much smaller than 14
that of the p-PAAm gels. As a result, the amount of energy dissipation and toughness of the p-15
PAAm gels are much larger than the PA gels. 16
17
Conclusions 18
In conclusion, we found that phase separation induces gel-glasslike transition of the 19
polyacrylamide gels equilibrated in DMF/water mixture. Around the transition point, the gel 20
shows very high stiffness, toughness, and self-healing. Since the origin of phase separation 21
was poor affinity between polymer of gel network and solvent, this phenomenon should not 22
be specific for PAAm-DMF system but should be found in other gel-solvent systems in 23
principle. Although this time we used combination of hydrophilic polymer and organic 24
solvent DMF that is harmful, we believe it is possible to prepare other tough phase-separated 25
gels based on a combination of hydrophobic polymer and harmless solvents such as water, 26
Page 14
1.
13
which can be widely used for medical or industrial materials in aqueous environment. 1
2
3
Experimental Section 4
Materials: Acrylamide (AAm) (Junsei Chemical Co., Ltd.) as a monomer was recrystallized 5
from chloroform. N,N’-Methylenebis (acrylamide) (MBAA) (Tokyo Kasei Kogyo Co., Ltd.) 6
as a cross-linking agent was recrystallized from ethanol. 2-Oxoglutaric acid (α-keto) (Wako 7
Pure Chemical Industries, Ltd.) as a photo-initiator was used as received. N,N-8
dimethylformamide (DMF) (Wako Pure Chemical Industries, Ltd.) and polyethylene glycol 9
(Mw; 20,000) (PEG20k) (Wako Pure Chemical Industries, Ltd.) were used as received. 10
11
Preparation of polyacrylamide gels: PAAm gels were synthesized by free-radical 12
polymerization. 0.1 mol% of MBAA and 0.1 mol% of α-keto, in respective to AAm, were 13
added to 2 M AAm aqueous solution. Under an argon gas atmosphere, the solution was 14
poured into the space between two glass plates with thickness of 3 mm separated by silicone 15
rubber spacer with thickness of 2-7 mm. Photo-polymerization was carried out by 365 nm UV 16
irradiation for 8 h at 25°C. After the polymerization, the gel was immersed in a large amount 17
of water for a week to reach an equilibrium state and to wash away any residual chemicals. 18
19
Immersion of polyacrylamide gels: DMF/water mixtures were prepared by mixing DMF and 20
water with various weight ratios, and PEG20k aqueous solutions were prepared by dissolving 21
PEG20k powder in water in various concentrations. The PAAm gels were immersed in 22
various DMF/water mixtures or PEG20k aqueous solutions at room temperature for at least 7 23
days. During immersion, the solutions were exchanged 2 or 3 times to maintain the 24
concentration of the solutions. 25
26
Page 15
1.
14
Calculation of Polymer Volume Fraction: To calculate polymer volume fraction ϕ of the p- 1
and o-PAAm gels, the water-swollen PAAm gels were cut to the predetermined disk shape (15 2
mm in diameter = d0) with a cutting machine specially designed for soft materials (SDAP-3
100N, Dumbbell Co., Ltd.) at first. The gels were then immersed into DMF/water or PEG 4
solutions. After reaching equilibrium in the DMF/water mixture or PEG solutions, diameter of 5
the gels d was measured with caliper. The volume-swelling ratio of the immersed gels α was 6
determined under the assumption of isotropic swelling by, 7
α = (𝑑
𝑑0)
3
Then, their polymer volume fraction f is calculated by, 8
𝜙 = 𝜙0 ×1
𝛼
where ϕ0 is polymer volume fraction of water-swollen PAAm gels. We obtained ϕ0 = 0.031, 9
since initial water content of the water-swollen PAAm gels in this study is 95.6wt%, density 10
of PAAm is 1.437 ± 0.2 g cm-3
, and density of water is 1 g cm-3
.[23]
11
12
Tensile Test: Uniaxial tensile experiments were performed in air using a commercial testing 13
machine (Instron 5965 type universal testing system). Samples of thickness = 1.3-2.3 mm 14
were cut into the dumbbell shape standardized as JIS K6251-7 (test region: width = 2 mm, 15
length = 12 mm), with a cutting machine specially designed for soft materials (SDAP-100N, 16
Dumbbell Co., Ltd.). Both ends of a test piece were clamped and pulled apart at a constant 17
velocity, v = 100 mm min-1
. The fracture stress and the fracture strain were defined as the 18
nominal stress and strain at breaking point, respectively. Young’s modulus was defined as 19
slope of initial stress-strain curves. Work of extension at fracture Wextf was defined as the area 20
under the stress-strain curve. 21
For loading and unloading tensile test, samples were stretched at a velocity of 100 mm min-1
22
Page 16
1.
15
(strain rate = 8.33 min-1
) to strain ε = 5 at room temperature, then were moved back to initial 1
displacement immediately at the same velocity of stretching. The energy dissipation can be 2
estimated from the hysteresis area Uhys by, 3
𝑈ℎ𝑦𝑠 = ∫ (𝜎𝑙𝑜𝑎𝑑 − 𝜎𝑢𝑛𝑙𝑜𝑎𝑑)𝑑𝜀𝜀=5
𝜀=0
Where σload is the stress in loading and σunload is that in unloading. 4
The loading and unloading tensile test for evaluating self-healing of sacrificial bond was 5
performed to one sample repeatedly. Samples were stretched to strain ε = 3 at velocity 100 6
mm/min. After each loading and unloading test, the sample was immersed in the 7
corresponding DMF/water mixture at room temperature for a prescribed time (10 min or 24 8
hr) before the next loading and unloading test. To accelerate the self-healing rate, the samples 9
were also thermal annealed by immersing in the corresponding DMF/water solution of 75oC 10
for 5 min to recover the initial form of the sample, and then cooled down in the solution at 11
room temperature for 5 min. In the latter case, the virgin samples before stretching were also 12
annealed for 5 min at 75oC before testing and tested at room temperature for considering 13
annealing effect on mechanical property of the gels. 14
15
Pure Shear Test: To evaluate the toughness of p-PAAm, the fracture energy T was measured 16
according to the method previously reported with a pure shear experiments.[14,18,22]
The 17
measurement was performed using the testing machine. Samples of thickness b0 = 1.25-2.70 18
mm were cut into the strip shape (width a0 = 40 mm, height H = 40-60 mm), with a cutting 19
machine specially designed for soft materials (SDL-200, Dumbbell Co., Ltd.). Both ends of a 20
test piece were clamped and pulled apart at a constant velocity, v = 100 mm min-1
from their 21
initial distance (L0 = 8 mm) between the two clamps, while the lower clamp was fixed 22
(Figure S2a). The force (F)-stretching length (L) curves of the samples were recorded. Figure 23
S2c is the force-stretching length curve of the unnotched (solid line) and notched (broken 24
Page 17
1.
16
line) samples of 80p-PAAm. The grey area is W (Lc), the work done by the applied force to 1
the unnotched sample at the critical stretching distance Lc that the notched sample start to 2
propagate the crack. Lc was determined by video monitoring. The fracture energy is calculated 3
by the equation
4
𝑇 =𝑊(𝐿𝑐)
𝑎0 × 𝑏0=
1
𝑎0 × 𝑏0∫ 𝐹(𝐿)𝑑𝐿
𝐿𝑐
0
5
Rheological Test: Storage and loss moduli of the samples were determined using an ARES 6
rheometer (TA Instruments, Co., Ltd.). Samples of 1.3-2.3 mm thickness were cut to the disk 7
shape (diameter = 15 mm) with the cutting machine SDAP-100N. Top and bottom surfaces of 8
the disk sample was fixed between the two parallel plates of the rheometer with glue and 9
covered by an aluminum cup. Then the solution used for gel immersion was poured into the 10
cup to prevent drying of the samples during measurement. Temperature-sweep measurement 11
was performed with the temperature range of 0.1-80.1oC at constant shear strain amplitude of 12
0.1% and frequency of 10 Hz. 13
14
15
Supporting Information 16
Supporting Information is available from the Wiley Online Library or from the author. 17
18
Acknowledgements 19
This research was financially supported by a Grant-in-Aid for Scientific Research (S) (No. 20
124225006) from the Japan Society for the Promotion of Science (JSPS). The authors thank 21
Dr. Tomoyasu Aizawa (Hokkaido University) for his kind assistance for NMR measurement. 22
23
Page 18
1.
17
Received: ((will be filled in by the editorial staff)) 1
Revised: ((will be filled in by the editorial staff)) 2
Published online: ((will be filled in by the editorial staff)) 3
4
References 5
[1] A. Suzuki, M. Yamazaki, Y. Kobiki, J. Chem. Phys. 1996, 104, 1751 6
[2] Y. Hirokawa, H. Jinnai, Y. Nishikawa, T. Okamoto, T. Hashimoto, Macromolecules 1999, 7
32, 7093 8
[3] M. Shibayama, M. Morimoto, S. Nomura, Macromolecules 1994, 27, 5060 9
[4] G. J. Lake, A. G. Thomas, Proc. R. Soc. Lond., A, Math. Phys. 1967, 300, 108 10
[5] Y. Tanaka, K. Fukao and Y. Miyamoto, Eur. Phys. J. E 2000, 3, 395 11
[6] D. Bonn, H. Kellay, M. Prochnow, K. Ben-Djemiaa, J. Meunier, Science 1998, 280, 265 12
[7] T. Masuike, S. Taki, K. Hara, S. Kai, J. Appl. Phys. 1995, 34, 4997 13
[8] P. G. de Gennes, Langmuir 1996, 12, 4497 14
[9] A. Coniglio, H. E. Stanley, W. Klein, Phys. Rev. Lett. 1979, 42, 518 15
[10] K. L. Spiller, S. J. Laurencin, A. M. Lowman, J. Biomed. Mater. Res. 2009, 90B, 752 16
[11] S. Wu., R. A. Shanks, J. Appl. Polym. Sci. 2004, 93, 1493 17
[12] Y.-H. Na, Y. Tanaka, Y. Kawauchi, H. Furukawa, T. Sumiyoshi, J. P. Gong, Y. Osada, 18
Macromolecules 2006, 39, 4641 19
[13] L. R. G. Treloar, The Physics of Rubber Elasticity, Oxford University Press, UK, 1975. 20
[14] R. S. Rivlin, A. G. Thomas, J. Polym. Sci. 1953, 10, 291 21
[15] J. P. Gong, Y. Katsuyama, T. Kurokawa, Y. Osada, Adv. Mater. 2003, 15, 155 22
[16] M. Huang, H. Furukawa, Y. Tanaka, T. Nakajima, Y. Osada, J. P. Gong, 23
Macromolecules, 2007, 40, 6658. 24
[17] T. Nakajima, H. Furukawa, Y. Tanaka, T. Kurokawa, Y. Osada, J. P. Gong, 25
Macromolecules, 2009, 42, 2184. 26
Page 19
1.
18
[18] J.-Y. Sun, X. Zhao, W. R. K. Illeperuma, O. Chaudhuri, K. H. Oh, D. J. Mooney, J. J. 1
Vlassak, Z. Suo, Nature 2012, 489, 133. 2
[19] J. Li , W. R. K. Illeperuma, Z. Suo, J. J. Vlassak, ACS Macro Lett. 2014, 3, 520 3
[20] J. Li, Z. Suo, J. J. Vlassak, J. Mater. Chem. B 2014, 2, 6708 4
[21] J.P. Gong, Soft Matter 2010, 6, 2583. 5
[22] T. L. Sun, T, Kurokawa, S, Kuroda, A. B. Ihsan, T. Akasaki, K. Sato, M. A. Haque, T. 6
Nakajima, J. P. Gong, Nature Mater. 2013, 12, 932 7
[23] T. Tominaga, V. R. Tirumala, S. Lee, E. K. Lin, J. P. Gong, W. L. Wu, J. Phys. Chem. B 8
2008, 112, 3903 9
[24] H. E. Gottlieb, V. Kotlyar, A, Nudelman, J. Org. Chem. 1997, 62, 7512 10
11
Page 20
1.
19
Figure Captions 1
Figure 1. Pictures and polymer volume change of PAAm hydrogels in DMF/water solvent 2
and PEG aqueous solution. The PAAm gels in high DMF concentration CDMF became turbid 3
(a) while o-PAAm gels kept transparent even at the same ϕ in 40 wt% of PEG solution (b). 4
Polymer volume fraction ϕ of p-PAAm gels increased abruptly at CDMF = 60wt% (c), while 5
that of o-PAAm gels monotonously increases with increasing PEG concentration CPEG (d) to 6
the same level of volume fraction with p-PAAm. Each measurement was performed for 3~8 7
times, and the errors of some measurements were so small that the error bars were hidden 8
behind the plots. 9
10
11
12
0
0.1
0.2
0.3
0.4
0.5
0.6
0 10 20 30 40
CPEG
(wt%)
0
0.1
0.2
0.3
0.4
0.5
0.6
0 20 40 60 80 100
CDMF
(wt%)
CDMF
(wt%)
0
a)
40 60 70 80 100 90
CPEG
(wt%) 0 20 30 40 10
b)
10 mm
10 mm
d) c)
Page 21
1.
20
Figure 2. Tensile performance of the p- and o-PAAm gels. (a, b) Stress-strain curves for p- 1
(a) and o-PAAm (b) gels under uniaxial tensile test. The results of turbid samples with phase 2
separation are represented by open symbols in (a). For clarification of relationship between 3
lines and sample, 1-10% of the data are shown as plots; (c, d) Young’s modulus E, fracture 4
strain εf, of p- and o-PAAm gels. The Young’s modulus and fracture strain of o-PAAm gels 5
monotonously increased with increasing CPEG (c), while the modulus abruptly increased at 6
CDMF = 60wt% and fracture strain abruptly dropped at CDMF = 80wt% (d). (e)Young’s 7
modulus of p- and o-PAAm gels against the PAAm volume fraction ϕ. Each mechanical test 8
was performed for 3~4 times, and the errors of some tests were so small that the error bars 9
were hidden behind the plots. 10
11
12
10-3
10-2
10-1
100
101
102
0 10 20 30 40 50
Ef
E (
MP
a)
f (m
m/m
m)
CPEG
(wt%)
d) c)
10-3
10-2
10-1
100
101
102
103
10-2
10-1
100
101
102
0 20 40 60 80 100
E
fE
(M
Pa)
f (m
m/m
m)
CDMF
(wt%)
b) a)
0
1
2
3
4
5
6
7
0 4 8 12 16
0wt% 0.03160wt% 0.2470wt% 0.4080wt% 0.4390wt% 0.40
Str
ess
(M
Pa)
Strain (mm/mm)
CDMF
0
0.1
0.2
0.3
0.4
0.5
0.6
0 5 10 15 20 25
0wt% 0.03120wt% 0.2045wt% 0.50
Str
ess
(M
Pa
)
Strain (mm/mm)
CPEG
Page 22
1.
21
1
2
e)
10-3
10-2
10-1
100
101
102
103
0.01 0.1 1
DMF
PEG
Yo
un
g's
Mo
du
lus (
MP
a)
E ~ 0.98
Page 23
1.
22
Figure 3. Fracture energy T of the p- and o-PAAm gels measured by pure-shear test. (a) T of 1
p-PAAm gels at various CDMF; (b) T of o-PAAm gels at various CPEG. Each test was 2
performed for 3 times, and the errors of some tests were so small that the error bars were 3
hidden behind the plots. To demonstrate the high strength and toughness of the 80p-PAAm 4
gel, the Japanese children game “kusa-sumo” (c) is performed, and the result of kusa-sumo 5
between a silicone rubber (upper light sample) and the 80p-PAAm gel (lower dark sample) is 6
shown (d). The gel was dyed by colored pen after setting. 7
8
9
10
c)
Initial state Gel won!
d)
leaf/stem
Silicone
Gel
a) b)
101
102
103
104
105
0 20 40 60 80 100
T (
J m
-2)
CDMF
(wt%)
101
102
103
104
105
0 10 20 30 40 50
T (
J m
-2)
CPEG
(wt%)
Page 24
1.
23
Figure 4. Energy dissipation of p- and o-PAAm gels. (a, b) Loading and unloading uniaxial 1
test results of p-PAAm gels of CDMF = 0-70wt% (a), and CDMF = 80wt% (b). Stretching 2
velocity: 100 mm min-1
. To make the figure clear, only 0.5% of the collected data are shown 3
in (a). (c, d) Work of extension at fracture Wextf and mechanical hysteresis loss Uhys of o-4
PAAm gels (c) and p-PAAm gels (d) at a strain range of 0 ≤ ε ≤ 5. Wextf and Uhys of o-PAAm 5
gels monotonously increased with increasing CPEG (c), while the Wextf showed a peak at CDMF 6
= 80wt% and Uhys showed maximum at the same CDMF (d). The results of turbid samples with 7
phase separation are represented by open symbols. Each loading and unloading tensile test 8
was performed for 1 time. (e, f) Repeated loading and unloading tensile test of non-thermal 9
annnealed (e) and thermal annealed (f) 80p-PAAm gels. For non-thermal anealed sample, the 10
waiting time between virgin and 2nd
, 2nd
and 3rd
, 3rd
and 4th
were 10 min, 10 min, 24 h, 11
respectively (e). For thermal anealed sample, the sample was anealed at 75 o
C for 5min. 12
between each test. p-PAAm recovers its original properties after 24 hours waiting from the 13
first test (e). Annealing can accererate this healing process, enabling to recover initial 14
properties within 5 min. (f). 15
16
0
1
2
3
4
5
6
7
0 1 2 3 4 5 6
80wt% 0.43
Str
ess (
MP
a)
Strain (mm/mm)
CDMF
b) a)
0
0.05
0.1
0.15
0 1 2 3 4 5 6
0wt% 0.03160wt% 0.2570wt% 0.42S
tress
(M
Pa
)
Strain (mm/mm)
CDMF
Page 25
1.
24
1
2
3
4
0
1
2
3
4
0 0.5 1 1.5 2 2.5 3
75oC -- Thermal Annealed
Virgin2nd-5min3rd-5min
Str
ess
(M
Pa)
Strain (mm/mm)
Virgin
2nd3rd
e) f)
0
1
2
3
4
5
6
7
0 0.5 1 1.5 2 2.5 3
25oC
Virgin2nd-10min3rd-10min4th-24h
Str
ess
(M
Pa)
Strain (mm/mm)
Virgin
2nd
3rd
4th
c) d)
10-2
10-1
100
101
102
10-3
10-2
10-1
100
101
102
0 20 40 60 80 100
Wextf
Uhys
We
xtf (
MJ m
-3) U
hy
s (MJ m
-3)
CDMF
(wt%)
10-2
10-1
100
101
102
10-3
10-2
10-1
100
101
102
0 10 20 30 40 50
Wextf
Uhys
Wex
tf (
MJ m
-3) U
hys (M
J m
-3)
CPEG
(wt%)
Page 26
1.
25
Figure 5. Illustration of deformation and fracture mechanism in phase-separated p-PAAm 1
gels. (a) In the case of low CDMF, polymer chains are highly solvated so that polymer-polymer 2
interaction in gels is screened. Gels obey classical rubber elasticity with negligible hysteresis 3
loss. (b) In the case of middle CDMF, polymer-polymer associations can be preferentially 4
destroyed during deformation, which gives a high toughness and self-healing of the material. 5
(c) In the case of high CDMF, gels show brittle and hard nature due to too strong polymer-6
polymer associations. 7
8
9
10
11
12
13
Deformation
c)
b)
Deformation
Deformation
a)